Mantle Plumes

The Earth's mantle is a solid rocky layer 3000km thick extending from the bottom of the crust (30km depth) to the core-mantle boundary (CMB) at the top of the hot molten iron core. The crust and the top 100km of the mantle form a stiff cold layer called the lithosphere. Underneath the lithosphere the mantle behaves like a viscous fluid. Over geological timescales of 10's to 100's of millions of years, the mantle convects. At the surface the lithosphere sinks into the mantle at subduction zones and spreads apart at mid-ocean ridges. At these ridges the mantle partially melts and basalt lava about 6km thick forms the new ocean crust. At the CMB, mantle material is heated by the core and hot mantle "plumes" rise through the mantle as large spherical "heads" (1000km in diameter) with thinner "tails" connecting the heads to the source region. When plume heads approach the surface they partially melt, leading to massive, geologically brief outpourings of lava. These "flood basalt" events occur about every 10-20 million years. The aim of this project is to simulate mantle plumes impinging on the bottom of the lithosphere and melting.

Principal Investigator

Geoff Davies
Geophysical Fluid Dynamics
Research School of Earth Sciences
Australian National University



Facilities Used



Alison Leitch
Department of Earth Sciences
Memorial University of Newfoundland

RFCD Codes


Significant Achievements, Anticipated Outcomes and Future Work

We studied the melting that occurs when a plume first approaches the bottom of the lithosphere, aiming to match the observed melt rates and volumes of flood basalts. Melt rates are very sensitive to temperature, depth of melting and composition. We found that to match the observation using reasonable plume temperatures and lithosphere thicknesses, the plume must contain a significant fraction (about 15%) of easily melted recycled ocean crust. This is consistent with the quantities of basalt that are expected to have been returned to the mantle during subduction throughout the 4.5 billion years of Earth's history. We also found that in order for the plume to reach shallow depths beneath the lithosphere it was important to include a realistic model of mantle viscosity. At a depth of 660km, the mantle viscosity increases (going down) by about an order of magnitude. As plumes encounter this viscosity step they neck down, and the small central, hottest part of the plume rises quickly to the bottom of the lithosphere. This behaviour is necessary to produce the very rapid voluminous melting that is geologically observed.

On-going work, which has involved significant code development, involves the impingement of plumes on lithosphere of varying thickness. Seismic investigations indicate that the bottom of the lithosphere is not smooth under continents, and ponding of plume material in the shallowest parts of the lithosphere could lead to localized melting. We are also studying the conduction of heat through the lithosphere from cooling, spreading mantle plume heads.


Computational Techniques Used

The simulations have been carried out using the finite-difference program CONMG, written by G.F. Davies and modified by A.M. Leitch. The code was optimized to run on the VPP and is very efficient. This is important because very high resolution is required to properly resolve the melting region in the plumes (10's of km) within a mantle-scale (3000km) simulation. The high core memory of the VPP allowed high resolution simulations to be carried out that could not be run on workstations. Lower resolution exploratory simulations run in tens of minutes rather than hours or days.


Publications, Awards and External Funding

A. M. Leitch, G. F. Davies, Mantle Plumes and Flood Basalts: Enhanced Melting from Plume Ascent and an Eclogite Component, Journal of Geophysical Research, 106, 2001, 2047-2060.

Work is partially funded through NSERC research operating grant of A.M. Leitch.